cGAS Mediates Inflammation by Polarizing Macrophages to M1 Phenotype via the mTORC1 Pathway

Cyclic GMP-AMP (cGAMP) synthase (cGAS) is a cytosolic DNA sensor that directly binds with DNA and catalyzes cGAMP production (1, 2). cGAMP further activates stimulator of IFN genes (STING) and subsequently promotes activation of the TBK1-IRF3 pathway and NF-κB pathway, leading to the production of type I IFN and other inflammatory cytokines, such as TNF-α and IL-6 (3–5). In general, cGAS-mediated inflammatory response is a pivotal defense against pathogenic infection, whereas aberrant overactivation of the cGAS-STING pathway frequently occurs and further induces severe inflammation and pathological tissue damage (6–9). Thus, cGAS signaling must be tightly controlled to restrict its effect in reasonable ranges. However, the exact role of cGAS in the inflammatory response and inflammation-involved diseases remains to be fully explored. Macrophages play a key role in the defense against pathogens and are also critically involved in the inflammatory response. Macrophages have high diversity and phenotypic plasticity in response to environmental signals (10, 11). Plasticity is a hallmark of macrophages, and the macrophages are able to be polarized to the proinflammatory M1 phenotype or the alternative M2 phenotype with distinct physiological functions (12). The proinflammatory M1 phenotype could be induced by IFN-γ and LPS, which further exert the proinflammatory effect via secreting numerous inflammatory factors, such as inducible NO synthase (iNOS), IL-6, and TNF-α (13–16). The anti-inflammatory M2 phenotype is triggered by IL-4 or IL-13 and characterized by high expression of mannose receptor (also called CD206), arginase 1 (Arg1), and the chitinase-like proteins including Ym1 and Fizz1 (16–18).

Macrophage plasticity is critical for the initiation and resolution of the inflammatory response, and cGAS is highly expressed in macrophages and actively involved in inflammation, which elicits the intriguing question about the role of cGAS in the process of macrophage polarization.

In this study, we investigated the regulatory effect of cGAS on macrophage polarization during inflammation. Our data showed that during infectious inflammation, LPS induced the release of mitochondrial DNA (mtDNA) into the cytosol, leading to the activation of cGAS-mTORC1 signaling and macrophage polarization to M1 phenotype. In addition, we demonstrated that cGAS knockout attenuated the acute inflammatory response in the cellular model as well as the animal model. Thus, this study paved a new avenue for the regulation of macrophage polarization and provided, to our knowledge, a novel strategy to manipulate the inflammatory diseases by targeting cGAS.

Cgas^−/− C57BL/6J mice were purchased from the Nanjing Biomedical Research Institute of Nanjing University. Tlr4^−/− C57BL/6J mice were purchased from The Jackson Laboratory. The animal model of acute lung injury was established by i.p. injecting 50 mg/kg LPS to mice for further investigation. All experiments with mice were performed in accordance with the general guidelines of the Institutional Animal Care and Use Committee. All performances related to mice were approved by the Ethics Committee of Shandong University. The lung wet/dry weight ratio was measured to evaluate the extent of pulmonary edema according to Li et al. (19).

LPS (L4130-100MG) was purchased from Sigma-Aldrich (Merck, Darmstadt, Germany). Mouse IFN-γ (315-05-20UG) and IL-4 (250-07-5UG) were purchased from PeproTech (Cranbury, NJ). Rapamycin (S1039) was purchased from Selleck Chemicals (Houston, TX). The anti-Arg1 Ab (sc271430, RRID: AB_10648473) was purchased from Santa Cruz (Dallas, TX). The Abs, including anti-cGAS (31659, RRID: AB_2799008), anti–phospho-p38 (4511, RRID: AB_2139682), anti–phospho-ERK (4370, RRID: AB_2315112), anti-p38 (9212, RRID: AB_330713), anti-ERK (4695, RRID: AB_390779), anti–phospho-STAT1 (7649, RRID: AB_10950970), anti–phospho-STAT3 (9145, RRID: AB_2491009), anti-p65 (8242, RRID: AB_10859369), anti–phospho-p65 (3033, RRID: AB_331284), anti–phospho-JAK2 (3771, RRID: AB_330403), anti–phospho-S6 (4858, RRID: AB_916156), and anti–phospho-4E-BP1 (2855, RRID: AB_560835), were purchased from Cell Signaling Technology (Beverly, MA). Anti-GAPDH (E-AB-20059, RRID: AB_2905551) Ab was purchased from Elabscience Biotechnology (Wuhan, China), and anti–β-actin (66009-1-Ig, RRID: AB_2883475) Ab was purchased from Proteintech (Chicago, IL). Anti-F4/80-PerCP/Cy5.5 (123128, RRID: AB_893484), anti-CD11b-FITC (101205, RRID: AB_312788), and anti-Ly6G-allophycocyanin (127613, RRID: AB_1877163) Abs were purchased from BioLegend (San Diego, CA). The anti-CD16/32-allophycocyanin (E-AB-F0997E) and anti-CD206-PE (E-AB-F1135D) Abs were purchased from Elabscience Biotechnology (Wuhan, China).

Male C57BL/6J mice (6–8 wk old) were i.p. injected with 1 ml of sterile starch to recruit peritoneal macrophages. Seventy-two hours after the injection, the recruited macrophages were collected into DMEM from the lavage of the peritoneal cavity. The preparation of bone marrow–derived macrophages (BMDMs) was performed as previously described (20). The isolated macrophages were treated with LPS (100 ng/ml) and IFN-γ (50 ng/ml) for 24 h for induction to the M1 phenotype. The macrophages were treated with IL-4 (20 ng/ml) for induction to the M2 phenotype. The alveolar macrophages were harvested from the bronchoalveolar lavage fluid (BALF) of mice according to Meng et al. (21).

The level of NO in macrophages was detected according to the procedure of the NO detection kit (S0021S, Beyotime Biotechnology, Shanghai, China). For the detection of 2′,3′-cGAMP, 2 × 10⁶ macrophages were treated with LPS (100 ng/ml) and IFN-γ (50 ng/ml) for 24 h before being harvested by M-PER protein extraction solution (Thermo Fisher Scientific, Waltham, MA). The level of 2′,3′-cGAMP was further analyzed according to the instructions of the 2′,3′-cGAMP ELISA kit (501700, Cayman Chemical, Ann Arbor, MI).

Immunohistochemistry, Western blot, quantitative PCR (qPCR), and ELISA were performed as previously described (20, 22). The primers were synthesized (Sango Biotech, Shanghai, China). The forward and reverse sequences of the primers can be provided upon request. The ELISA assay was performed to detect mouse IL-6 according to the manufacturer’s instructions (E-EL-M0044c, Elabscience Biotechnology, Wuhan, China). The flow cytometry assay was performed (Beckman Coulter, Indianapolis, IN) and further analyzed by the CytExpert software (Beckman Coulter, Indianapolis, IN).

Extraction and detection of mtDNA were performed as previously described (23). qPCR was performed by CFX96 (Bio-Rad Laboratories, Hercules, CA). Primers for mitochondrial cytochrome c oxidase subunit I (mtCOI) were as follows: forward, 5′-GCCCCAGATATAGCATTCCC-3′; reverse: 5′-GTTCATCCTGTTCCTGCTCC-3′. Primers for the 18S RNA were as follows: forward, 5′-TAGAGGGACAAGTGGCGTTC-3′; reverse, 5′-CGCTGAGCCAGTCAGTGT-3′.

Data were presented as mean ± SD and statistically analyzed using GraphPad Prism software (version 9.0, GraphPad Software, La Jolla, CA). Statistical significance among groups was evaluated with a two-tailed Student t test or two-way ANOVA. A p value of <0.05 was considered statistically significant.

All relevant data that support the findings of this study are available from the corresponding author upon request.

Macrophage polarization plays a pivotal role in inflammation, and cGAS, as an innate immune sensor highly expressed in macrophages (https://www.immgen.org), is also recognized to play a critical role in the inflammatory response. Thus, this elicits an intriguing question about the role of cGAS in macrophage polarization, which we try to elucidate in this study.

We collected macrophages from mice and further detected the expression level of cGAS under the induction of macrophage polarization. The data showed that cGAS expression was significantly upregulated under the stimulation of LPS and IFN-γ (Fig. 1A), and it was downregulated under the stimulation of IL-4 (Fig. 1B), which indicated the involvement of cGAS in macrophage polarization. The data further revealed that the M1 markers were significantly suppressed (Fig. 1C, Supplemental Fig. 1A), whereas the M2 markers were significantly upregulated in macrophages isolated from the Cgas knockout (Cgas^−/−) mice (Fig. 1D). Macrophages can produce NO by iNOS in response to LPS challenge, and our data revealed that the production of the recognized M1 marker NO was also reduced in the Cgas^−/− macrophages (Fig. 1E). Both the data from the mouse peritoneal macrophages and BMDMs showed that the protein level of Arg1, a marker of M2 macrophages, was significantly upregulated in the Cgas^−/− macrophages (Fig. 1F). In addition, macrophage polarization to the M1 phenotype was inhibited after Cgas was deleted (Supplemental Fig. 1B, 1C). Further investigation showed that both the mRNA and protein levels of cGAS in macrophages were significantly upregulated in a dose- and time-dependent manner in response to LPS stimulation (Supplemental Fig. 1D–F). Analysis of inflammatory signaling showed that activation of the inflammatory MAPK and NF-κB pathway, as well as the production of inflammatory cytokines, including TNF-α and IL-6, was significantly inhibited in Cgas^−/− macrophages (Fig. 1G–I). Altogether, these data indicated that cGAS was able to enhance the inflammatory response by promoting macrophage polarization to the M1 phenotype.

It is recognized that mtDNA is released into the cytosol in stressed cells (24), and thus we are interested in defining whether the cytosolic mtDNA activates cGAS during inflammation. It is known that mtCOI was encoded by mtDNA, and its level represents the amount of cytosolic mtDNA. Thus, the level of mtCOI was quantitatively detected by qPCR, and the data showed that the amount of cytosolic mtDNA in the LPS-stimulated macrophages was significantly increased (Fig. 2A). Further investigation showed that the expression of cGAS in macrophages was significantly upregulated under stimulation of mtDNA or LPS and IFN-γ (Fig. 2B). The Western blot data further revealed that mtDNA significantly increased the phosphorylation level of STAT1 and STAT3, whereas this effect was almost completely abrogated by the deletion of Cgas (Fig. 2C), indicating that mtDNA exerted its effect via activation of cGAS. Further investigation showed that M1 markers were significantly upregulated under the stimulation of mtDNA (Fig. 2D), indicating that mtDNA promoted macrophage polarization to the M1 phenotype. Altogether, these data indicated that during inflammation, cGAS was activated by cytosolic mtDNA and further promoted macrophage polarization to the M1 phenotype.

As TLR4 is the receptor of LPS, and the recent literature revealed that the TLR-mediated cGAS signaling was involved in the immune response against HIV infection (25), we further tried to define whether TLR4 signaling is involved in cGAS-mediated macrophage polarization. Our data showed that the amount of cytosolic mtDNA in LPS-stimulated Tlr4^−/− macrophages was much smaller compared with their wild-type (WT) counterpart (Fig. 2E). Under stimulation of LPS and IFN-γ, the upregulated cGAS expression was abrogated in Tlr4^−/− macrophages (Supplemental Fig. 2A, 2B); additionally, the production of cGAMP was also significantly attenuated after deletion of Tlr4 (Fig. 2F), indicating that LPS-TLR4 signaling participated in the regulation of cGAS activation in macrophages. Further investigation showed that M1 markers were significantly downregulated in Tlr4^−/− macrophages (Supplemental Fig. 2C), but induction to the M2 phenotype was not affected (Supplemental Fig. 2D), indicating that TLR4 played an important role in the regulation of cGAS-mediated macrophage polarization to the M1 phenotype.

cGAS is an IFN-stimulated gene and it also mediates IFN-β production, leading to the formation of a positive feedback loop when cGAS is activated (26). Thus, we further tried to define whether there was a positive feedback loop involved in cGAS-mediated macrophage polarization. The data showed that the IFN-β level was elevated in M1 macrophages, which in turn induced macrophage transformation to the M1 phenotype and upregulated cGAS expression, suggesting that there was a positive feedback loop in cGAS-mediated macrophage polarization (Supplemental Fig. 2E, 2F). We further transfected cGAMP, the product of activated cGAS, into the Cgas^−/− macrophages before induction of macrophage polarization. The data showed that cGAMP could not completely rescue polarization to the M1 phenotype in the Cgas^−/− macrophages (Supplemental Fig. 2G), which suggested the involvement of other signaling in cGAS-mediated macrophage polarization other than the classical cGAS-STING pathway.

To further explore the mechanism involved in the cGAS-mediated macrophage polarization to the M1 phenotype, we induced macrophage polarization in vitro and tested the possible involved pathways. The data showed that when macrophages were induced to the M1 phenotype, activation of the mTORC1 pathway was significantly enhanced (Fig. 3A), whereas its upregulation was abrogated in the Cgas^−/− macrophages and the Tlr4^−/− macrophages (Fig. 3B, 3C), indicating the possible involvement of mTORC1 signaling in the cGAS-induced macrophage polarization. Further data showed that activation of the mTORC1 pathway was significantly enhanced under the stimulation of mtDNA (Fig. 3D), whereas both LPS/IFN-γ– and mtDNA-induced mTORC1 pathway activation was significantly abrogated in Cgas^−/− macrophages (Fig. 3E), indicating the pivotal role of cGAS in the mtDNA-activated mTORC1 pathway.

When we blocked the mTORC1 pathway by its specific inhibitor rapamycin, the LPS/IFN-γ–stimulated upregulation of the M1 markers was significantly inhibited in WT macrophages, whereas Cgas deletion significantly abrogated this effect (Fig. 3F, 3G). In contrast, blocking the mTORC1 pathway increased the expression of M2 markers in WT macrophages, whereas it had no significant influence on Cgas^−/− macrophages (Fig. 3H). Altogether, these data indicated that cGAS induced macrophage polarization to the M1 phenotype through the mTORC1 pathway.

It is recognized that macrophage polarization critically affects sepsis-related inflammation and organ injury, and thus we further tried to define the role of cGAS in sepsis-related systemic inflammation and organ failure. We constructed a mouse acute inflammation model by i.p. injecting LPS. The mice significantly slowed down, and a series of sepsis-related symptoms appeared 6 h after the injection, such as shortness of breath, elevated body temperature, and tremors, indicating the successful construction of the animal model. We further estimated the inflammatory status of these mice, which revealed that the proportions of macrophages and neutrophils in the peripheral blood and spleen of Cgas^−/− mice were both significantly decreased compared with those of the WT mice (Fig. 4A, Supplemental Fig. 3A, 3B). Further investigation showed that the overall survival time of LPS-challenged Cgas^−/− mice was significantly prolonged compared with that of WT mice (Supplemental Fig. 3C). Thus, these data indicated that deletion of Cgas significantly inhibited acute inflammation and attenuated systemic inflammatory damage in mice.

We analyzed the expression of cGAS in mouse organs and it showed that the lung tissue had the highest expression of cGAS (Fig. 4B). Further analysis from the HPA, GTEx, and FANTOM5 databases also verified the high expression of cGAS in the lung tissue (Supplemental Fig. S3), and the cGAS expression in the alveolar macrophages is high (ImmGen database). After construction of the LPS-induced sepsis mouse model, the cGAS expression was significantly upregulated in the lungs of the LPS-challenged mice (Fig. 4C), whereas the tissue damage, especially in the lungs, was significantly alleviated in Cgas^−/− mice (Fig. 4D, Supplemental Fig. 3E).

The ratio of lung weight to body weight was also significantly decreased after the deletion of Cgas (Fig. 4E), indicating a significant recovery of the Cgas^−/− mice from the sepsis-induced acute lung injury. Altogether, these data indicated that cGAS aggravated sepsis-induced tissue damage in mice, and Cgas deletion significantly ameliorated pulmonary injury during sepsis.

Accumulating evidence suggests that macrophage polarization critically affects sepsis-related inflammation and organ injury, including acute lung injury (27), and thus we further tried to define whether cGAS deletion alleviates acute lung injury via its regulation of macrophage polarization. The qPCR data showed that the inflammatory cytokines in the lung tissues, including Tnfa and Il6, as well as the macrophage chemokine Ccl2 and the neutrophil chemokine Cxcl2, were all significantly decreased in the Cgas^−/− mice compared with the WT mice (Fig. 5A). Further flow cytometry assays showed that the amount of CD11b⁺ and F4/80⁺ macrophages in the BALF of Cgas^−/− mice was also significantly decreased (Fig. 5B). These data indicated a significantly decreased inflammatory response and reduced macrophage recruitment in the lung tissues of Cgas^−/− mice compared with WT mice.

The data further showed that when Cgas was deleted, the proportion of infiltrated M1 macrophages in the BALF was significantly decreased, whereas the proportion of M2 macrophages was significantly increased (Fig. 5C, 5D). The amount of M1 macrophages infiltrated in the lung tissue was significantly decreased, whereas the amount of infiltrated M2 macrophages was significantly increased in Cgas^−/− mice compared with WT mice (Fig. 5E, 5F), confirming that Cgas deletion attenuated the lung injury via its regulation of macrophage polarization. Further investigation verified that Cgas deletion induced macrophage polarization through its regulation of the mTORC1 pathway (Fig. 5G), which verified that cGAS induced macrophage polarization to the M1 phenotype via its activation of the mTORC1 pathway. Collectively, these findings indicated that cGAS significantly increased the proportion of M1 macrophages in BALF, and Cgas deletion significantly ameliorated sepsis-induced acute lung injury.

Altogether, this study demonstrated that cGAS promoted the conversion of macrophages to the inflammatory M1 phenotype through the mtDNA-mTORC1 pathway, and Cgas deletion significantly ameliorated sepsis-induced acute lung injury (Fig. 5H).

The newly discovered DNA sensor cGAS is activated by cytosolic DNA and further mediates innate immune responses against pathogens (28). Although cGAS is recognized to be actively involved in the inflammatory response (29), whether it has any effect on macrophage polarization remains to be fully clarified. In this study, we identified cGAS as a macrophage polarization switch and demonstrated that in the inflammatory response, the mtDNA was released into the cytosol via the LPS-TLR4 pathway, which subsequently led to the activation of cGAS signaling and macrophage polarization to the proinflammatory M1 phenotype.

In healthy cells, host DNA normally resides in the nucleus or mitochondria. Whereas in certain pathophysiological conditions, the mitochondrial homeostasis is broken under oxidative stress, and mtDNA is released into the cytosol where it serves as the danger-associated molecular pattern and further triggers the cGAS-mediated immune response (30). LPS is the toxic component of Gram-negative bacteria, and it is responsible for the pathogenesis in most infectious inflammation (31). In this study, we demonstrated that under the challenge of LPS, mtDNA was released into the cytosol and further activated cGAS signaling, leading to macrophage polarization to the M1 phenotype. In addition, LPS significantly induced cGAS expression via the TLR4 pathway, with the outcome of inflammation aggravation. Consistent with our study, the regulation of cGAS by LPS was reported in A549 lung cancer cells (32), and the activation of cGAS-STING signaling through the TLR4 pathway was involved in inflammasome activation (29). However, to our knowledge, the regulatory effect of cGAS on macrophage polarization is defined in the present study for the first time.

In this study, we demonstrated that cGAS promoted macrophage polarization to the M1 phenotype via activation of the mTORC1 pathway. mTOR, as an evolutionarily conserved serine/threonine protein kinase, functions to orchestrate diverse environmental signals and translate these cues into appropriate cellular responses, including innate immune response and inflammation (33–35). Thus, it is noteworthy that the signaling of cGAS is regulated by mTOR complexes and vice versa, although the underlying mechanism remains to be clarified (36, 37). Many studies suggest that cGAS activation is modulated by environmental inputs, and the response of immune cells, such as macrophages, could be tailored and optimized according to specific environmental necessities, indicating the possible regulatory role of cGAS in macrophage polarization. It is recognized that the mTORC1 pathway plays a role in the plasticity of mononuclear phagocytes (33–35), whereas, to our knowledge, the cGAS-mediated regulation of the mTORC1 pathway and macrophage polarization was investigated in the present study for the first time. We demonstrated that cGAS induced macrophage polarization to the M1 phenotype via the mTORC1 pathway, which was rescued by inhibiting mTORC1 signaling by rapamycin. Thus, our results indicated a pivotal role of the mTORC1 pathway in cGAS-induced macrophage polarization, and they also suggested the possibility of regulating cGAS-mediated inflammatory diseases by rapamycin.

LPS is the main toxic component of bacteria responsible for the pathologic damage to the host via inducing destructive inflammation, and thus LPS-induced sepsis is commonly used to mimic the fulminant systemic inflammation-mediated multiple tissue damage and organ failure (29, 38). Based on the high expression of cGAS in lung tissue, we are interested in defining the role of cGAS in acute lung injury, a common clinical disease that results from critically severe sepsis. It has been recognized that macrophage polarization from the M1 to M2 phenotype significantly attenuates the inflammatory lung injury (39), and thus we further tried to define the role of cGAS-mediated macrophage polarization in sepsis-induced acute lung injury. Our data revealed that knockout of Cgas significantly alleviated lung injury and improved its physiological condition in the sepsis mouse model. Analysis of the infiltrated immune cells in BALF of the sepsis mouse model also verified that cGAS mediated the macrophage polarization to switch from the M1 to M2 phenotype. Thus, these data demonstrated that cGAS promoted macrophage polarization to the M1 phenotype and its knockout alleviated sepsis-induced acute lung injury via regulating macrophage polarization, which further indicated, to our knowledge, a novel manipulation strategy for fatal sepsis by modulating cGAS.

In conclusion, in the present study, we demonstrated that in infectious inflammation, cGAS was activated by mtDNA in the cytosol of macrophages, and it further activated the mTORC1 pathway, promoted macrophage polarization to the M1 phenotype, and aggravated acute lung injury in the LPS-induced sepsis mouse model. Thus, we revealed the role of cGAS-mediated inflammation from a new perspective and indicated a potential treatment of clinical inflammatory diseases by targeting cGAS.

The authors have no financial conflicts of interest.

We thank the Translational Medicine Core Facility of Shandong University for consultation and instrument availability that supported this work.